Autonomous competitive research in minutes. Analyzes websites, captures ad screenshots from LinkedIn/Meta/Google ad libraries, and outputs structured reports.
Claude Code CLI
~/.claude/agents/
Browserbase.com
claude "run..."
# Install Claude Code CLI
npm install -g @anthropic-ai/claude-code
# Create agents folder & save the .md file
mkdir -p ~/.claude/agents
mv ~/Downloads/competitive-intel.md ~/.claude/agents/
# Install puppeteer for screenshots
npm install -g puppeteer-core
# Run the agent
claude "run competitive-intel agent for acme.com"
Uses Browserbase cloud browsers to capture full-page screenshots from ad libraries.
// Create cloud browser session
const session = await fetch('https://www.browserbase.com/v1/sessions', {
method: 'POST',
headers: { 'x-bb-api-key': BROWSERBASE_API_KEY },
body: JSON.stringify({ projectId: BROWSERBASE_PROJECT_ID })
}).then(r => r.json());
// Connect puppeteer to cloud browser
const browser = await puppeteer.connect({
browserWSEndpoint: `wss://connect.browserbase.com?apiKey=${BROWSERBASE_API_KEY}&sessionId=${session.id}`
});
// Capture screenshot from LinkedIn Ad Library
await page.goto('https://www.linkedin.com/ad-library/search?accountOwner=acme');
await page.screenshot({ path: 'acme-linkedin-ads.png', fullPage: true });
Save the .md file to ~/.claude/agents/ and start running competitive research.